National disidentification and minority identity: A study among Muslims in Western Europe

Abstract

Host national disidentification in which immigrants explicitly distance themselves from society is problematic for a cohesive national community and is likely to hamper immigrants’ successful host society integration. Among Sunni Muslim immigrants of Turkish origin living in Germany and the Netherlands we tested whether (a) an empirical distinction between national disidentification and identification can be made, (b) whether higher perceived group discrimination of Muslims (“reactive religiosity path”) and (c) stronger Muslim self-centrality (“intrinsic religiosity path”) are associated with stronger host national disidentification, to the extent that they are associated with a stronger commitment to religious identity content. Disidentification was found to be a separate construct and both the reactive religiosity path and the intrinsic religiosity path were found to be related to higher levels of disidentification, mediated by religious identity.

Keywords:

• Disidentification
• discrimination
• religious identity content
• immigrants

In contrast to the United States where religion used to be seen as facilitating immigrants’ integration process (Foner & Alba, 2008), immigrant religion (particularly Islam) is viewed as a problematic area in West European countries that are more secular (Zolberg & Long, 1999). Islam forms a “bright boundary” (Alba, 2005) separating immigrants from host societies and is at the heart of what in Europe is considered a “crisis of multiculturalism” (Modood, 2007). In many European countries and increasingly also in the United States (Pew, 2015), Muslim identity and host national identity are polarized. Islam has become the “negative other” and symbolic for problems related to immigrants and immigration. Anti-Muslim sentiments are widespread in Western countries (Arab American Institute, 2015; Strabac & Listhaug, 2008) and more widespread than anti-foreign resentments (Spruyt & Elchardus, 2012; Verkuyten, Hindriks, & Coenders, 2016). Leading politicians in Europe and the United States have publicly described Islam as a “problem” and a “backward religion” that seriously threatens society (see Verkuyten, 2014), and political parties have incorporated in their party program that “Islam does not belong to the host nation” (e.g., “Alternative for Germany”; “Party for Freedom” in the Netherlands). In addition, research in Western Europe has found that a majority of Muslim immigrants considers themselves primarily a Muslim rather than a national of the host country (Pew 2006; Phalet & Güngör, 2004; Verkuyten & Martinovic, 2012).

In the present study we examine national disidentification among Turkish Muslim immigrants in Germany and the Netherlands, as a crucial, but largely neglected, aspect of psychological distancing from the host society. The aim is to examine perceived “extrinsic” and “intrinsic” factors that might be related to individual differences in national disidentification, namely (a) perceived group discrimination of Muslims, and (b) religious self-centrality. Furthermore, we study whether these factors are related to religious identity content. Disidentification refers to the active rejection and distancing of the national category and we argue that this is more likely when the content of the one (Muslim) identity is contradictory to that of the (national) other. We use the conceptualization proposed by Kellstedt, Green, Guth, & Smidt (1996) to assess three aspects of religious identity content – belief, belonging and behavior – as potential mediators in the relation between perceived group discrimination and religious self-centrality with national disidentification.

National disidentification

Disidentification, or distancing from the national identity is qualitatively different from low national identification which implies a low sense of national belonging and little commitment to the nation. Low identification means that aspects of the national category are not strongly connected to oneself, but this does not have to consist of disconnecting these aspects from oneself (Kreiner & Ashforth, 2004). Studies in organizational contexts have shown that disidentification is a different psychological state than low identification (e.g., Elsbach & Bhattacharya, 2001; Kreiner & Ashforth, 2004). Disidentification is not merely a lack of identification, but rather implies an adversarial stance in which a group membership is rejected (“not wanting to belong”). It involves minority group members developing a reactive or oppositional identity in which individuals actively separate their minority identity from the culture and defining aspects of the dominant group (Fuller-Rowell & Doan, 2010; Ogbu, 1993; but see Ainsworth-Darnell & Downey, 1998).

Yet, research on immigrant and minority groups has interpreted a low score on a scale of national identification as indicating disidentification (e.g., Jasinskaja-Lahti, Liebkind, & Solheim, 2009). However, disagreeing with an item such as “I identify with the host nation” does not have to indicate disidentification in which minority members distance themselves from society and the majority group is considered “what we are not” or the “not-us” (McCall, 2003). Research also has used questions on momentarily not feeling like a member of a certain group to assess disidentification (Yip, 2016). Yet, indicating that at a particular moment of the day one does not feel like a group member does not have to indicate disidentification with that group.

To our knowledge, almost no research has empirically examined this form of disidentification among minority groups. An exception is a research among immigrants in the Netherlands that confirmed the empirical distinction between host national identification (e.g., “I identify with the Dutch”) and disidentification (e.g., “I certainly do not want to see myself as Dutch”) (Verkuyten & Yildiz, 2007). In addition to trying to replicate this finding, we build on this research by assessing factors explaining disidentification among Muslim migrants. More specifically, we consider the role of perceived Muslim discrimination as an “extrinsic” source for host national disidentification, and of religious self-centrality as an “intrinsic” source for disidentification.

Perceived religious group discrimination

Group discrimination implies unjust disadvantages and unfair treatment of one’s minority group which tells people that they are not equal members of society (Tyler, 2001). Discrimination presents a threat to the value of one’s group identity and members of devalued groups can cope with identity threat by distancing themselves from the majority group (Richman & Leary, 2009). Disidentification can be used as a defensive mechanism in an effort to maintain a positive sense of self (Elsbach & Bhattacharya, 2001). Thus, minority members can distance themselves from the national category when they feel that their minority group is devalued and excluded in society. Some studies have found suggestive evidence for this (e.g., de Vroome, Verkuyten, & Martinovic, 2014). For example, studying ethnic minority youth in Belgium, Sweden and The Netherlands, Fleischmann (2011) found that perceived personal discrimination was associated with weaker national identification, and in France (e.g., Badea, Jetten, Iyer, & Er-rafiy, 2011) and in the United States (Sirin & Fine, 2008) similar results have been found. In a longitudinal study among Russian immigrants in Finland, it was found that perceived personal discrimination was causally related to lower national identification (Jasinskaja-Lahti et al., 2009; see also Mähönen, Jasinskaja-Lahti, & Liebkind, 2011). Furthermore, not only perceived personal discrimination but also perceived group discrimination has been found to be independently associated with immigrants’ weaker orientation towards the host society (e.g., Fleischmann & Verkuyten, 2016; Ten Teije, Coenders, & Verkuyten, 2013; Verkuyten & Yildiz, 2007), and ethnic minority members’ national identification (Molina, Phillips, & Sidanius, 2015). The awareness that one’s minority group is discriminated can be associated with a lower sense of community belonging (Elliott & Doane, 2015), especially when there is a relatively strong minority group identification. However, these studies did not examine disidentification as distancing from the national identity. We expect that perceived group discrimination results in higher levels of disidentification among Muslim immigrants in Germany and the Netherlands.

One of the key mechanisms through which discrimination can result in higher levels of national disidentification, is via increased commitment to one’s religious identity. Research in the social identity tradition has observed a positive relationship between perceptions of discrimination and in-group commitment (see Schmitt & Branscombe, 2002). This so-called rejection-identification model argues that social rejection and devaluation causes minorities to increasingly turn towards their minority group and as a result more strongly distance themselves from the majority group (Branscombe, Schmitt, & Harvey, 1999). This process of rejection and minority group identification has been found among many groups, including African Americans (Branscombe et al., 1999), Mexican Americans (Molina et al., 2015; Romero & Roberts, 2003), atheists (Doane & Elliott, 2015), and Muslims (Sirin & Fine, 2008) in the United States. Existing European research among Muslim immigrants also suggests that group discrimination is positively associated with religious group identity (Fleischmann, Phalet, Deboosere, & Neels, 2012; Karlsen & Nazroo, 2013; but see Maliepaard & Gijsberts, 2012), but whether this also leads to disidentification is yet to be determined. Perceived group discrimination makes group boundaries salient and relatively impermeable and thereby leads to a feeling of identity conflict in which the normative beliefs and expectations of one’s religious minority identity are considered to be incompatible with a sense of host national belonging (Hirsh & Kang, 2015). Thus, we expect higher perceived group discrimination to be associated with stronger national disidentification, through a stronger commitment to Islam (the “reactive religiosity path”).

Religious self-centrality

Whereas group discrimination can be seen as a perceived extrinsic factor affecting disidentification, Muslim migrants also have their own religious norms, beliefs and practices which might contradict a sense of host national belonging. Muslim group identity is not only a form of “reactive religiosity” resulting from perceived group discrimination, but also based on early socialization and learning as a form of “intrinsic religiosity” (Myers, 1996; de Hoon & Van Tubergen, 2014; Verkuyten, Thijs, & Stevens, 2012). Research has shown intergenerational religious transmission among Muslim immigrant families (e.g., Güngör, Fleischmann, & Phalet, 2011) and that for Muslims living in Western Europe, religion has great importance in the way they live their lives (Haddad & Smith, 2001; Vertovec & Rogers, 1999). Nevertheless, there are individual differences in religious beliefs and practices (Roy, 2007). We argue that in addition to discrimination, religious self-centrality plays a role in shaping religious identity content, which in turn affects national disidentification. Religious self-centrality is expected to be related to disidentification to the extent that it involves religious identity content which is difficult to reconcile with norms, values and practices of the host society (the “intrinsic religiosity path”).

Religious identity content

In their organizing framework for conceptualizing social identity, Ashmore, Deaux, and McLaughlin-Volpe (2004) make a distinction between social identity aspects such as group-based beliefs, interdependence, and behavioral involvement. Similarly, for conceptualizing and measuring religious identity Kellstedt and colleagues (1996) propose the three-B classification of religious belief, belonging and behavior. These three aspects have been found to have different and sometimes conflicting effects on, for example, political and social tolerance (Ben-Nun Bloom & Arikan, 2012; Nunn, Crockett, & Williams, 1978). We will develop arguments as to how these three different aspects might be related to national disidentification.

Fundamentalist beliefs

Religious belief is concerned with the moral good and divine truth which is difficult to reconcile with moral and epistemic diversity (Ysseldyk, Matheson, & Anisman, 2010). This is particularly the case when beliefs are fundamental in nature. Fundamentalist believers are strict in their attitude to religious texts (Almond, Appleby, & Silvan, 2003) and fundamentalism can be defined as “The belief that there is one set of religious teachings that clearly contains the fundamental, basic, intrinsic, essential, inerrant truth about humanity and deity” (Altemeyer & Hunsberger, 1992, p. 118). This definition implies religious norms and ideals that emphasize a single, unchangeable interpretation which is binding for all believers.¹

According to the religious value conflict model (Brandt & Van Tongeren, 2015) the more emphasis a religion places on fundamentalist beliefs, the less tolerant the followers will be towards dissimilar groups that (implicitly) challenge or violate these beliefs (Beatty & Walter, 1984; Froese, Bader, & Smith, 2008; Tutiya, 2005). It has been demonstrated that people reject those with dissimilar beliefs in order to protect the validity and vitality of their own belief system (Brandt & Van Tongeren, 2015). Western countries have a long Christian history, have become increasingly secular (Bruce, 2011) and are questioning the presence of Islam. In line with this, preliminary evidence indicates that Muslims in Europe who hold more fundamentalist beliefs have more negative out-group attitudes (Koopmans, 2015) and are more inclined to support the political organization of Muslims (Verkuyten, 2016). Therefore, we expected stronger fundamentalist belief to be associated with stronger national disidentification.

Belonging to the Ummah

A second important aspect of social identity is group belonging which implies interdependence and mutual fate perceptions (Ashmore et al., 2004). In Islam there is an emphasis on Muslims forming a single community of believers (“Ummah”). A common Islamic community bound by its religion and patterned after the community founded by Muhammad is central to the faith. For people with a strong sense of belonging to their religious community, it is the functioning and fate of the community as a whole that gains importance, relative to one’s fate as an individual. It is this aspect of religious identity that is particularly likely to translate group membership into an in-group based orientation. The feeling among Muslim minority members that their fate is linked makes them more likely to feel an inner obligation to stick together and to define themselves in opposition to the majority group (e.g., Sanchez, 2006; Stokes-Brown, 2003).

Behavioral involvement

A third aspect of religious identity is behavioral involvement. Religious activities are forms of identity enactment which create a stronger and more visible boundary between the religious in-group and other groups. Previous studies show that people who regularly partake in religious practices (behavior) tend to be less accepting of religious outgroups than those who are not involved in such practices (e.g., Sullivan, Piereson, & Marcus, 1982). The more emphasis a religion places on strict religious rules and practices, the less accepting the followers will be towards groups that do not follow these rules and practices (Beatty & Walter, 1984; Froese et al., 2008; Tutiya, 2005). Islam is a religion with a strong emphasis on orthopraxis, presenting precise rules, practices and guidelines for living in accordance with the will of Allah. These rules and practices are the most visible markers of Islamic faith and can symbolize the incompatibility of Muslim life with western values and norms, as is illustrated by the headscarf controversy and other debates in European countries about the “not fitting in” of (alleged) Islamic rules and practices (Verkuyten, 2014). Therefore, it can be expected that the more Muslim immigrants engage in religious practices (e.g., Ramadan, praying five times a day) the more likely they are to reject the host society. Thus a positive association between behavioral involvement and host national disidentification was expected.

The current research

To summarize, our aim is to examine the relatively novel and important construct of national disidentification among Turkish Muslim immigrants in Germany and the Netherlands. We expected higher perceived group discrimination of the religious group (“reactive religiosity path”) and stronger Muslim self-centrality (“intrinsic religiosity path”) to be associated with stronger national disidentification, to the extent that they are associated with a stronger commitment to the three aspects of religious identity content (belief, belonging, and behavior; see Figure 1).

Figure 1. Structural model with unstandardized effects and standard errors between brackets: only significant paths are represented.

Given that the concept of disidentification is new in research on immigrants and minorities, it is important to ascertain that it is not only theoretically but also empirically different from low national identification. We will therefore test the discriminant validity of the two concepts by examining first whether the constructs of national identification and disidentification can be distinguished empirically, using confirmatory factor analysis. Second, we will examine whether the factors predicting identification and disidentification differ by estimating path models for disidentification and national identification simultaneously. Furthermore, we will take possible differences between Germany and the Netherlands into account, and we will explore whether the expected associations are similar or different in both countries.

Data and methods

In Germany and the Netherlands data collection procedures were similar. Participants were recruited by means of snowball sampling, with four chains being initiated in cities in each country.² Four Turkish research assistants initiated the data collection by visiting mosques and non-religious Turkish cultural centers, and by using their own networks. In this way we tried to reach participants that differ in their religious engagement. In Germany participants (46%) were mainly recruited in the western region around Cologne and in the Netherlands (54%) in the center (Utrecht) and east (Arnhem) of the country. They were asked to participate in a study on “Living in Germany/The Netherlands.” Questions were asked either in Dutch or German. These two languages are quite similar and the back-translation procedure (Dutch to German) was used to ensure comparability of questions. For the purposes of our study, Sunni respondents (N = 464) were selected.³

Men were somewhat overrepresented in the sample (57.6%). The age of the participants ranged from 16 to 80 years (M = 36.5, SD = 12.6). The mean length of residence in the host country was 22.13 years (SD = 9.99), and length of residence was positively associated with age (r = .49, p < .001). A majority of respondents (79.8%) had Dutch or German nationality, including those with a double nationality (Turkish-German or Turkish-Dutch). Regarding education, all levels were represented in the data, with 31% of participants having no formal degree or only a primary school diploma, while 40% was in a possession of different types of secondary school diploma, and 29% held a tertiary degree. The number of missings was very low (less than 2% per item).

Measures

National disidentification was measured with five items (7-point scales) taken from previous research (Verkuyten & Yildiz, 2007). These items were “I would never say ‘we Dutch/Germans’,” “I certainly do not want to see myself as Dutch/German,” “I always have the tendency to distance myself from the Dutch/Germans,” “Actually, I do not want to have anything to do with the Dutch/Germans,” and “I never feel addressed when people talk about the Dutch/Germans” (α = .77).

National identification was measured with five items (7-point scales) that are commonly used in social psychological research and that were taken from previous research on the national identification of immigrants (Verkuyten & Martinovic, 2012): e.g., “I feel Dutch/German” and “I feel connected to the Dutch/Germans” (α = .86).

Perceived group discrimination of Muslims in Germany/the Netherlands was assessed with four items (7-point scales): e.g., “Muslims are often discriminated against in daily life”; “Muslims are often discriminated against when trying to find work/an internship” (α = .84).

Religious self-centrality was measured by the following three items (7-point scales) taken from previous research in the Netherlands (Verkuyten & Yildiz, 2007): “Being Muslim is the most important thing in my life,” “Being Muslim is the only thing that really counts in my life,” and “Being Muslim determines how I feel inside.” Strong item wordings were used because previous findings have found “total religious group identification” (mode of 7) among Turkish Muslims in Western Europe when using general identification items (e.g., Verkuyten, 2007; Verkuyten & Yildiz, 2007). Even with this strong wording, the means of all three items are significantly (p_s < .001) above the midpoint of the scale (α = .90).

As for the three mediators: Fundamentalist belief was measured by the items “There is only one truth, which is the divine truth,” and “Islam has a fixed and unchangeable meaning” (r = .58). Belonging to the Ummah was measured by the two items “I feel very connected to the Ummah” and “my faith is connected to that of other Muslims” (r = .63). Behavioral involvement is a composite scale of participation in Ramadan, frequency of prayer and frequency of Mosque attendance (α = .78).

Given that our model includes three aspects of religious identity content, it is important to establish that these aspects are empirically distinct. Confirmatory factor analysis indicated that a single factor model fits the data poorly χ² (14, 462) = 232.738; χ²/df = 16.62, p > .05; CFI = .88, RMSEA = .18, and a two-factor model fits somewhat better, Δχ² = 51.693 with 1 df, p < .001. However, the three-factor structure based on the theoretical constructs fits the data best, Δχ² = 81.286 with 2 df, p < .001; χ² (10, 462) = 70.437; χ²/df = 7.0, p > .05; CFI = .96, RMSEA = .10.

Analysis

In the structural models, we controlled statistically for various demographic variables which could relate to both our dependent and independent variables and thus influence our findings if omitted: country of residence (1 = The Netherlands, 0 = Germany), gender (1 = female, 0 = male), age, education – a scale ranging from (1) no education to (9) university degree, host country nationality (1 = yes, 0 = no), and length of stay in the host country in years.

Concerning the statistical power, the sample size is adequate (N > 200 is “large” in SEM models, Kline, 2005) and well above the 10 cases per variable rule of thumb. We have 51 free parameters in our model which exceeds Bentler and Chou’s (1987) free parameter-sample size ratio. In addition, we used bootstrapping to estimate the indirect relations in order to increase the power in the model (see Fritz & MacKinnon, 2007). However, considering the sample size and the high number of variables (37) and paths to be estimated (such as item loadings to the latent constructs, direct and indirect paths, paths from control variables on mediator and outcome variables), we opted for a manifest (rather than latent) model in the estimation of our structural path model (using bootstrapping). The manifest structural model (which included control variables) had an excellent fit χ² (9, 428) = 7.115; χ²/df = .79, p = .625; CFI = 1.00, RMSEA = .00.

Results

National (dis)identification

As expected, confirmatory factor analysis indicated that national disidentification and identification are separate constructs. A comparison of a single-factor and a two-factor model in Mplus shows that the single factor model fits the data very poorly, χ² (35, 463) = 722.246; χ²/df = 20.64, p > .05; CFI = .63, RMSEA = .21. The fit of the two-factor model is significantly better Δχ² = 406.77 with 1 df, p < .001. Furthermore, all standardized loadings on the respective factors are >.5, and there are no cross-loadings (<.15). Thus, although both factors are associated (r = −.56), host national disidentification is empirically distinct from identification. As shown in Table 1, the average agreement with disidentification is at the midpoint of the scale (t(444)=−1.011, p = .313, 95% CI [−.179; .057], 56% neutral or higher), and the agreement with the identification scale is significantly below the midpoint of the scale (t(444)=−5.357, p < .001, 95% CI [−.493, −.228]; 49% score neutral or higher).

Table 1. Descriptive statistics of the different constructs.

	Range	Mean	SD
National disidentification	1–7	3.94	1.27
National identification	1–7	3.62	1.41
Perceived discrimination	1–7	4.52	1.47
Self-centrality	1–7	4.31	1.82
Fundamentalist beliefs	1–7	4.87	1.67
Belonging to the Ummah	1–7	3.89	1.90
Behavioral involvement	1–5	3.18	1.27

Explaining national disidentification

As can be seen in Table 2 and Figure 1, Muslims with more fundamentalist beliefs show higher levels of disidentification with the host society. In addition, Muslims who more strongly feel part of the Muslim community of believers show higher levels of disidentification. There is no difference in disidentification between those Muslims who practice their religion frequently and those who do not.

Table 2. Full structural model predicting national disidentification and national identification.

	B (S.E.)	95% CI
National disidentification
Perceived discrimination	.086 (.050)	−.011, .185
Self-centrality	−.021 (.051)	−.120, .080
Fundamentalist beliefs	.194 (.053)***	.089, .297
Belonging to the Ummah	.099 (.041)*	.021, .181
Behavioral involvement	.081 (.056)	−.034, .190
National identification
Perceived discrimination	.072 (.061)	−.047, .189
Self-centrality	−.134 (.061)*	−.256, −.017
Fundamentalist beliefs	−.069 (.059)	−.182, .049
Belonging to the Ummah	.029 (.049)	−.072. .120
Behavioral involvement	−.086 (.068)	−.222, .045
Fundamentalist beliefs
Perceived discrimination	.153 (.059)**	.042, .270
Self-centrality	.505 (.043)***	.419, .586
Belonging to the Ummah
Perceived discrimination	.189 (.053)***	.089, .297
Self-centrality	.582 (.041)***	.499, .661
Behavioral involvement
Perceived discrimination	.033 (.040)	−.044, .112
Self-centrality	.412 (.041)***	.354, .467

Note: Controls not shown.

^* p < .05

^** p < .01

^*** p < .001.

Results indicate that perceived group discrimination is positively related to two of our three mediating factors: belonging to the Ummah and fundamentalist beliefs, but not to behavioral involvement. In line with theorizing, those who perceive more discrimination of Muslims in society feel a stronger sense of belonging to the Muslim community of believers and hold more fundamentalist beliefs. Results from our mediation analyses indicate that the indirect path from perceived discrimination to disidentification via fundamentalist belief is significant, B = .030, se = .013, p < .05; 95%, CI [.011, .064]. The indirect path via belonging to the Ummah is marginally significant, B = .019, se = .010, p = .054, 95% CI [.004, .042]).

When we turn to our second independent variable, stronger religious self-centrality is associated with stronger behavioral involvement, feelings of belonging to the Ummah and holding more fundamentalist beliefs. In turn, there are indirect effects on national disidentification through fundamentalist beliefs, B = .098, se = .029, p < .01; 95% CI [.044, .157], and belonging, B = .057, se = .025, p < .05; 95% CI [.013, .110], but again not through behavioral involvement, B = .033, se = .023, p > .1; 95% CI [−.014, .078].

When comparing standardized effects (not shown), the indirect effects of self-centrality are larger than those of perceived group discrimination. Both the direct path from perceived discrimination and the one from self-centrality to national disidentification become non-significant after including the mediating factors, indicating full mediation.

National identification

In order to assess whether the associations found for disidentification differ from those for identification, we simultaneously modeled national disidentification and identification. Figure 1 shows that national identification is not associated with fundamentalist beliefs, belonging, or behavioral involvement. Furthermore, neither perceived group discrimination nor religious self-centrality are related to national identification through any of the three proposed mediators. Finally, there is no direct effect of perceived group discrimination on national identification, while Muslims with stronger religious self-centrality show a somewhat lower national identification.

Additional analyses

In our theoretical model, both perceived group discrimination and religious self-centrality are antecedents of religious beliefs, belonging and behavior, representing an extrinsic and intrinsic path. From the rejection-identification model (Schmitt & Branscombe, 2002), however, it could be argued that religious self-centrality is wholly endogenous, and predicted by perceived discrimination, in the same way as the three mediators. Therefore, we compared the fit of this alternative model with our proposed model using the Aikaike Information Criterion (AIC). A lower AIC indicates a better fit, and while there is no formal significance test, differences larger than 10 are considered meaningful (Burnham & Anderson, 2004). Comparing the AICs of both models clearly shows that our theorized model (AIC = 6859) fits the data better than the alternative model (AIC = 8487; ΔAIC = 1628).⁴

Given the fact that we have data from two countries, we also investigated whether the associations between the various constructs are comparable across countries. Separate models for Germany and the Netherlands show that the findings are largely similar to those discussed above. The significant effects of self-centrality and perceived group discrimination on the mediators are replicated in each country; there are no direct effects of discrimination and self-centrality on disidentification in either country; and the positive effect of fundamentalist beliefs on disidentification is found in both countries. Yet, there are also some country differences. The effect of belonging to the Ummah on disidentification is positive in both countries but only reaches significance in Germany. In addition, in Germany there is an effect of behavioral involvement on disidentification, whereas this is not the case in the Netherlands, nor in the full sample.

Discussion

The current study is one of first to examine the novel and important construct of national disidentification among immigrants. The study makes a unique contribution to our understanding of what is involved in Muslim immigrants developing an adverse psychological orientation towards the host society. An increasing number of studies focuses on the question of immigrants’ sense of belonging and commitment to the nation (see Verkuyten & Martinovic, 2012). Yet, these studies tend not to consider the role of religious identity and do not focus on national disidentification. Our findings show that national disidentification is empirically distinct from national identification, and that religious identity is most clearly associated with disidentification. This indicates that it is important to go beyond examining the levels of minority group identification and national identification that are the focus of most of the existing research (e.g., Fleischmann & Verkuyten, 2016; Jasinskaja-Lahti et al., 2009). Identity conflict is more likely to express itself in the process of disidentification rather than in low group identification, as disidentification implies perceived incompatibilities between the values, beliefs and behaviors connected to different social groups to which people belong (Hirsh & Kang, 2015). Low national identification refers to a weak sense of national belonging and this does not have to indicate identity conflict, while disidentification involves an adversarial stance in which a host national group membership is rejected.

National disidentification was stronger among Muslims with a stronger religious identity, in particular for the identity aspects (Ashmore et al., 2004) of belonging to the Ummah and fundamentalist belief. Emphasizing as part of the Ummah implies a strong in-group orientation that can set Muslims apart from the host society. Additionally, Muslim immigrants who more strongly endorsed religious beliefs about the basic and intrinsic truths of the social and metaphysical world (Altemeyer & Hunsberger, 1992) indicated higher national disidentification. This is in line with the religious values conflict model (Brandt & Van Tongeren, 2015) which argues that stronger fundamentalist beliefs go together with stronger rejection of dissimilar beliefs that (implicitly) challenge one’s own belief system. West European countries have a long Christian history and have become increasingly secular (Bruce, 2011) and an emphasis on the Christian nature of the national identity is stronger in European countries with more non-Christian immigrants (Kunovich, 2006).

Unexpectedly, behavioral involvement, as a third aspect of religious identity content (Kellstedt et al., 1996), was not independently associated with national disidentification. While religious practices can be regarded as a form of identity enactment creating boundaries between groups, involvement in these practices was not associated with taking an adversarial stance toward the host society. A possible explanation is that religious practices can have various other meanings than intergroup ones. Individuals can, for example, engage in these practice for personal religious reasons (e.g., praying) or for socializing with fellow believers (e.g., participating in Ramadan, Mosque attendance). It is also possible that religious practices have somewhat different meanings in different national contexts. In Germany we did find a significant association between stronger involvement in religious practices and higher national disidentification. In terms of granting citizenship rights, Germany is one of the most restrictive countries in Europe and the Netherlands one of the least restrictive (Koopmans, Michalowski, & Waibel, 2012). This might mean that engagement in Islamic practices has a stronger boundary drawing function in Germany compared to the Netherlands.

We further examined perceived group discrimination and religious self-centrality as a “reactive religiosity path” and an “intrinsic religiosity path,” respectively. In line with our theoretical expectations, we found empirical evidence for both these paths. Perceived discrimination and religious self-centrality were associated with Muslim identity, and through fundamentalist beliefs and belonging to the Ummah, indirectly with national disidentification. Research on minority group identity often focuses only on the important role of perceived discrimination, but minority identity (particularly religious minority identity) also depends strongly on socialization within the family and in-group community (Chafetz & Ebaugh, 2000; Güngör et al., 2011; Hughes et al., 2006). In line with this, our findings indicate that not only the (negative) societal context but also a developed sense of religious self-centrality is relevant for national disidentification. It is however important to note that religious self-centrality is only associated to disidentification to the extent that it is related to belonging to the Ummah and fundamentalist beliefs, and that religious self-centrality is thus not inherently or automatically associated with national disidentification. Furthermore, the associations with disidentification were generally not very strong (see also Verkuyten & Yildiz, 2007). This indicates that even strong fundamentalist beliefs and feelings of belonging to the Ummah should not be equated with a self-evident distancing from the host society.

Limitations

We had theoretical reasons for testing the proposed model and alternative models fitted the data less well. Hence, the current findings show important and theoretically informed directions of influence on national disidentification of Muslim immigrants in Germany and the Netherlands. Furthermore, there is experimental (Molina et al., 2015) and longitudinal (Jasinskaja-Lahti et al., 2009) evidence for the impact of perceived discrimination on ingroup identification. Nevertheless, we cannot rule out that a sense of belonging to the Ummah might lead to perceptions of group-based discrimination and stimulates religious self-centrality. It is possible that there are mutual influences including the ones that we tested and longitudinal research is needed to further examine potential bidirectional influences.

Another limitation of the research is that it is not possible to draw a random sample of Sunni Muslim immigrants (because of the lack of official registration) and therefore a snowball sample was used. This limits the possibility to draw generalizable conclusions. We tried to reach a group of people that differs in their religious engagement by recruiting participants not only through the Mosques but also through cultural organizations and personal networks. Yet, some form of selectivity is likely, also because the questions were asked in the host languages. Although the great majority of Turkish immigrants is able to speak and read these languages, some subgroups (first generation, older women) are not (e.g., Huijnk & Dagevos, 2012). This probably means that we did not reach the most segregated Turkish Muslims. Though this can have implications for the mean levels of the variables, we don’t expect it to influence the associations between the measures. Furthermore, the quite similar findings in the two countries suggest that we have presented a more general picture about Muslim immigrants of Turkish origin in Western Europe. It seems that the processes involved are not strongly affected by country differences in opportunity structure and the institutional incorporation of Muslims (Statham & Tillie, 2016), and that more general psychological processes are involved in national disidentification of Muslim immigrants. Still, future studies should investigate this further by examining these issues in other countries and among other Muslim immigrant (sub)groups to further improve our understanding of the ways in which psychological processes involved in national disidentification are shaped by the (national) context.

We focused on Sunni Muslim but national disidentification could also be examined among Shiite and Alevi Muslim immigrants. There are for example important differences in religious beliefs and practices between Sunnis and Alevis. Alevi Muslims tend to interpret Islam and the Qur’an in a spiritual and mystical way rather than in terms of strict rules and regulations. For most of them, the love of God and how a person treats other humans and whether he or she acts as a responsible and caring human being is considered a key issue. Some argue that “Alevis share with Germans and Europeans a democratic, laicist and egalitarian outlook” (in Kosnick, 2004, p. 985). Given this humanistic and democratic outlook, Alevi Muslim immigrants might be less inclined to disidentify with western host societies and a stronger religious identity might not be associated with stronger disidentification.

Furthermore, it is important for future studies to consider the content of the national identity. Self-defined multicultural and immigrant countries such as Canada and the United States offer more opportunities for immigrants and minorities to develop a sense of host national belonging than non-settler European countries with a long history of an established and dominant majority group such as Germany and the Netherlands. In addition to the religious identity content, the national identity content thus seems to be an important factor to consider.

Conclusion

National disidentification can be considered problematic for the stability and cohesion of society and has psychological implications. While it can be useful in defining a distinctive and positive minority identity (Elsbach & Bhattacharya, 2001), it might also have negative consequences for sociocultural adjustment and societal participation (Ogbu, 1993). Some argue that an emphasis on Muslim identity does not have to imply a lack of host national commitment, while others claim that both identities are incompatible (see Verkuyten, 2014). The current findings indicate that stronger Muslim identity is associated with a stronger distancing from the host society. An important reason for this identity conflict is the group-based discrimination that Muslim immigrants face and that leads to a stronger emphasis on their Muslim identity. Another reason is the self-centrality that one’s religion has with the related normative beliefs and standards that can be difficult to reconcile with the liberal and secular principles of the host society.

This pattern of findings raises the important question of how host national disidentification can be minimized while Muslim religiosity is maintained. It is clear that anti-Muslim sentiments and processes of discrimination and exclusion should be addressed to prevent Muslim immigrants from feeling like second class citizens. Additionally, some have argued for the need to develop an “Euro-Islam” or “Europeanized Islam” (Tibi, 2002) whereas others argue for the importance of the further recognition and institutionalization of Islam in Western Europe (Halm, 2013). Furthermore, higher education, labor market integration and political participation can be expected to contribute to a stronger sense of host national belonging (de Vroome et al., 2014), without necessarily undermining one’s religious identity. Future studies on both the origins and consequences of religious identity and national disidentification, and studies among various ethnic and religious groups, including Muslim subgroups, should contribute to a further understanding of national disidentification processes in relation to the nature of the groups and the intergroup context.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. This does not include the endorsement of violent means to achieve fundamentalist aims because these aims can also be pursued in peaceful, democratic ways (Emerson & Hartman, 2006).

2. These data are safely stored at the faculty data-storage. Part of these data have been used in previous research papers but the current research question, focus and analysis is novel. Thus, the findings have not been reported elsewhere and the paper does not duplicate previously published work.

3. The original sample included a group of Alevi respondents. These were excluded from the analyses, given the fact that Alevi identity is defined in linguistic, cultural, political or religious terms by those who call themselves Alevi.

4. We also tested a second alternative model, in which disidentification predicts self-centrality, which in turn predicts belonging, beliefs and behavior, which then predict disidentification and identification. This model also fits the data less well than our theorized model (AIC = 8497; ΔAIC = 1638).